2 * TRANSPORT: inbound sessions are not properly PING-PONG validated,
3 connect-event is then not properly signalled (peer behind NAT
4 never gets a 'connect' message) and finally we DO transmit
5 data messages over the invalidated, only half-connect-signalled
6 channel (which confuses the core state machine, or now causes
7 transport-api assertion failures). [CG]
8 * FS-acceptance testing [CG]
18 - trust: need *fast* way to check/update trust in peers
19 (async peerinfo would not be right; certainly not with the
22 - scheduler should change OS process priority based on task priority;
23 should make better use of task priorities in general
24 - only connect() sockets that are ready (select()) [Nils]
25 [On W32, we need to select after calling socket before
28 - design & implement new setup tool
30 - good to have for transport/DV evaluation!
32 - write DV API (need to move declarations from dv_api.c to gnunet_dv_service.h!)
33 - implement DV service
34 - implement DV library (looks done)
35 - implement DV transport plugin
37 - implement performance tests
39 - does not seem to work with timeouts (especially if service is not running)
41 - needs more testing (especially F2F topology)
42 - needs to re-try connecting after disconnect (currently, it
43 initially triggers a connection request, but if that connection
44 fails / goes down, it does not retry in a timely fashion;
45 cause seems to be the 'blacklist_after_attempt' being set to 1h,
46 which is rather long -- and should probably be adjusted based on
47 the number of connections / known peers)
48 - If the topology daemon crashes, peers that were put on the
49 blacklist with transport will never be removed from it (until
50 transport service dies); we should use the blacklist notification
51 API to learn about the exact set of blacklisted peers at all times
52 (FIXME: the transport_api implementation of blacklisting
53 also does not work nicely for this since it won't let us know about
54 disconnect-reconnect events and the implicit whitelisting
55 that might happen here; that's not so bad since we will
56 re-blacklist on pre-connect attempts anyway, so this is
57 a minor issue; OTOH, we might want to be more explicit about
58 allowing/forbidding connects on pre-connect to avoid
59 entering connect attempts to just be blacklisted shortly afterwards).
60 - the code uses the term 'blacklist' for both peers that are forbidden
61 to connect (i.e. F2F mode) as well as peers that we currently
62 won't try to actively connect to ourselves (since we just tried);
63 This is confusing. We need two distinct terms (greylist?).
64 - move code to use hash table instead of linked list
65 - instead of periodically discarding blacklisted entries,
66 simply add task that is triggered at the right time (earlier free,
68 - check if new HELLO learned is different from old HELLO
69 before resetting entire state!
71 - support recursive download even if filename is NULL and we hence
72 do not generate files on disk (use temp_filename)
73 - bound parallelism (# fs downloads)
74 - distinguish in performance tracking and event signalling between
75 downloads that are actually running and those that are merely in the queue
76 - persistence support (publish, unindex, search, download)
77 - active migration support (in fs or in datastore or new daemon?)
78 - gnunet-service-fs (hot-path routing, load-based routing, nitpicks)
79 - [gnunet-service-fs.c:208]: member 'LocalGetContext::results_bf_size' is never used
80 - [gnunet-service-fs.c:501]: member 'PendingRequest::used_pids_size' is never used
81 - [gnunet-service-fs.c:654]: member 'ConnectedPeer::last_client_replies' is never used
82 - [gnunet-service-fs.c:669]: member 'ConnectedPeer::avg_delay' is never used
83 - [gnunet-service-fs.c:675]: member 'ConnectedPeer::avg_priority' is never used
84 - [gnunet-service-fs.c:688]: member 'ConnectedPeer::pending_requests' is never used
85 - [gnunet-service-fs.c:694]: member 'ConnectedPeer::last_p2p_replies_woff' is never used
86 - [gnunet-service-fs.c:700]: member 'ConnectedPeer::last_client_replies_woff' is never used
88 + active reply route caching design & implementation of service; gap extension!
90 - API lacks cancellation methods (needed? or is disconnect enough?);
91 may also want to integrate request queuing here instead of gnunet-service-fs_drq.c
93 - use g_main_context_set_poll_func to integrate GTK with GNUnet Scheduler!? (YUCK!)
94 - OR: add scheduler API to enable integration with GTK main loop instead of doing our own select
95 - use g_main_context_pending, g_main_context_query / g_main_context_check / g_main_context_dispatch
96 and NEVER g_main_loop_run (can this be done? might be the clean way to do this! But how
97 to integrate this with "tak_main"? Docu says:
98 "It's OK to use the GLib main loop directly instead of gtk_main(), though it involves
99 slightly more typing. See GMainLoop in the GLib documentation."
100 => so maybe it "just works"?
104 - good to have for DV/DHT evaluation!
106 - implement DHT service
107 - implement DHT library
108 - implement testcases
109 - implement performance tests
113 - need to get rid of synchronous API for service starts (cause all kinds of problems)
114 [=> eliminate for need to tell ARM about service starts most of the time!] [Safey]
115 - better tracking of which config changes actually need to cause process restarts by ARM.
116 - listen for requests to discover dependencies between services (and avoid
117 having to explicitly program start requests)
118 - better crash management (attach debugging support, capture and analyze
119 debug output, detect random vs. deterministic crashes)
122 - datastore reservation (publishing)
123 - search: availability probes
124 - location URIs (publish, search, download)
125 - non-anonymous FS service (needs DHT)
126 + DHT integration for search
127 + CS-DHT-functions (DHT-put of LOC)
128 + P2P-functions (DHT-get)
129 - collection API & tests
130 + gnunet-pseudonym (collection support)
131 - implement FS performance tests
138 * Determine RC bugs and fix those!
140 - modify configuration to allow controlling connections for non-local starts
141 - testbed creation with topology (needs working F2F topology)
143 - implement testcases for distributed testing
144 - test basic peer re-configure
145 - test topology creation
146 - test churn generation
147 - consider changing API for peer-group termination to
148 call continuation when done
150 - finalize API design
153 - integration with transport service
154 * MYSQL database backends: [CG]
160 - expand bibliography
161 - convert documentation pages to books
162 - update books (especially for developers)
163 - create good Drupal theme for GNUnet
164 - make a NICE download page and figure out how to
165 enable developers to publish TGZs nicely
166 - port "contact" page
167 - add content type for "todo" items?
168 * POSTGRES database backends: [CG]
171 * Determine RC bugs and fix those!
175 - SMTP transport backend
176 - HTTPS transport backend
177 + improved HTTPS support in MHD
179 - Implement method of learning our external addresses from
180 other peers; need some kind of threshold-based
181 scheme, limiting both the total number of addresses that we accept
182 this way as well as requiring multiple confirmations; also, we
183 should possibly try to confirm that the given address works for
184 us ourselves (loopback-style) before adding it to the list
185 + we may be able to simplify WELCOME messages (no need to add
186 addresses there anymore, but may help to learn them there anyway...).
187 + we probably want some kind of voting/counting for learning IP addresses
188 (maybe including IP addresses in ads proportional to how often others
189 report them? we at least need some protection against >64k HELLOs!),
190 + provide a way to give the user a list of "learned" IP addresses and
191 a way to easily "veto" addresses off the list!
192 => If MiM attacker uses vetoed address, blacklist the specific IP for
193 the presumed neighbour!
194 - implement gnunet-transport (transport configurator / tester)
195 - UPnP-based IP detection
196 (Note: build library always, build service when libxml2/etc. are available)
198 - Remove KBlocks in gnunet-unindex (see discussion with Kenneth Almquist on gnunet-devs in 9/2009)
199 - Allow checking of presence of search results and/or content via command-line tools
200 (add options to gnunet-search / gnunet-download to limit search to local peer)
202 - expire 'ancient' HELLOs (those without valid addresses AND that
203 we have not 'used' (for their public keys) in a while; need a way
204 to track actual 'use')
205 - make sue we also trigger notifications whenever HELLOs expire
212 - should use hash map to look up sessions
214 - api creates many, many short-lived TCP connections; either some
215 clients should use the API differently or we need to change the
216 API to enable re-use of connections to the service
218 - should use BIO instead of mmap
220 - need to periodically probe latency/transport cost changes & possibly switch transport
221 - instantly filter addresses from *other* peers that
222 are *equal* to our own address + port (i.e., localhost:2086). We
223 no longer filter those for outgoing (helps with loopback testing
224 and keeps the code clean), but we should filter strictly *impossible*
225 incoming addresses! This is for efficiency, not correctness.
226 - should use hash map to look up Neighbours
228 - 'server' uses 'GNUNET_PEERINFO_iterate', should probably switch to notification API
229 (for more instant / up-to-date hostlists at lower cost) [OPTIMIZATION]
231 - check for duplicates on insertion (currently, same content is frequently
232 stored again [seen with KBLOCKS and SBLOCKS]!)
236 - repeatedly resolve hostname and look up interfaces to determine our own IP
237 - [./transport/plugin_transport_tcp.c:391]: (style) struct or union member 'Plugin::address_update_task' is never used (related to issue above)
239 - [./transport/gnunet-service-transport.c:173]: (style) struct or union member 'TransportPlugin::rebuild' is never used (related to TCP not refreshing external addresses?)
241 - add stats (# bytes available, # bytes used, # PUTs, # GETs, # GETs satisfied)
243 - support inline data in directories for recursive file downloads (fs_download)